Interpretations are useful: penalizing explanations to align neural networks with prior knowledge. Authors:Laura Rieger, Chandan Singh, W. James Murdoch, Bin Yu.
確定! 回上一頁